How Netflix Runs Fewer Clusters Without Failing Requests 🚀 |
Автор: Perfology
Загружено: 2025-10-11
Просмотров: 681
When incoming traffic spikes beyond cluster capacity, some calls are bound to fail — but what if they didn’t have to fail randomly? 🤔
In this talk, we explore how Netflix designed a system to automatically tune, test, and deploy customized load shedding configurations for each gRPC cluster — allowing them to safely run fewer and smaller clusters while minimizing overload risk.
You’ll learn:
⚙️ How Netflix reasons about cluster buffer capacity
📊 Automation techniques to maintain optimal utilization
🚦 What worked (and what didn’t) for system utilization signals
🎯 How to prioritize requests and engage load shedding intelligently
#NetflixEngineering #LoadShedding #gRPC #Scalability #Microservices #DistributedSystems #SRE
Follow Us on Facebook: / perfology
Connect on Instagram: / perfologys
Network on LinkedIn: / perfology
Доступные форматы для скачивания:
Скачать видео mp4
-
Информация по загрузке:
![🔥Microservices Master Plan [Hindi]](https://ricktube.ru/thumbnail/sZl-q7zykpE/mqdefault.jpg)
